System for evaluating intervention effect of yinqiao shuangjie mixture based on patient disease course structure

Abstract

The present application relates to the technical field of case data mining, in particular to a Yinqiao Shuangjie mixture intervention effect evaluation system based on patient disease course structure. Through the collection of multi-source data of different syndromes by the table cold and heat meter, the remission factors of the initial table cold syndrome and the internal heat syndrome of the patient are analyzed according to the initial disease condition, the influence between different syndromes at different time points is analyzed combined with the syndrome change characteristics, and the remission factors are adjusted; based on the remission effect of Yinqiao Shuangjie on different syndromes combined with the adjustment remission factors of the syndromes, the disease course risk parameters are determined, the deep learning model is trained through the dynamic evolution of the disease course over time, and the evaluation model is obtained. The present application comprehensively measures the disease course risk characteristics through the remission dynamic changes of different syndromes of table cold and internal heat, combines the deep learning model, improves the accuracy of intervention effect evaluation, and provides data support for the determination of subsequent auxiliary drug decision-making.

Description

Technical Field

[0001] This invention relates to the field of case data mining technology, specifically to a system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure. Background Technology

[0002] Nonbacterial upper respiratory tract infections are among the most common diseases, and in Traditional Chinese Medicine (TCM), they are classified as "external cold and internal heat" syndromes. This syndrome is characterized by a complex interplay of external wind-cold symptoms (such as chills, lack of sweating, and headache) and internal lung-heat symptoms (such as sore throat, cough, and yellow phlegm). Yinqiao Shuangjie Compound, a representative TCM preparation for this syndrome, utilizes the wind-dispelling and heat-clearing properties of Forsythia suspensa and Lonicera japonica, combined with the lung-clearing and fire-purging properties of Scutellaria baicalensis and Syzygium aromaticum, to achieve a synergistic therapeutic effect.

[0003] Modern pharmacological studies have shown that the active ingredients in the Yinqiao herb pair, such as chlorogenic acid, forsythoside, and caffeoyl derivatives, possess multi-target biological effects, exerting therapeutic effects through various mechanisms including antiviral, anti-inflammatory, and immunomodulatory mechanisms. However, traditional Chinese medicine efficacy evaluation often relies on subjective symptom descriptions and experience-based judgments. As the disease progresses, external cold may rapidly transform into internal heat, turning into pure heat syndrome or heat excess damaging body fluids syndrome. Furthermore, the evolution of these syndromes may vary among different patients due to differences in their underlying conditions. This makes it difficult to obtain accurate and reliable data support from subjective experience or static data analysis, resulting in low accuracy of the evaluation results. Summary of the Invention

[0004] To address the technical problem in existing technologies where the disease progression and underlying conditions of different patients can lead to varying outcomes, and where subjective experience or static data analysis lacks accurate and reliable data support, resulting in low accuracy of evaluation results, the present invention aims to provide a system for evaluating the intervention effect of Yinqiao Shuangjie Compound based on the patient's disease progression structure. The specific technical solution adopted is as follows: This invention provides a system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure, the system comprising: The data acquisition module is used to obtain the integral vector of different syndrome types for each patient in the historical sampling database at each sampling time. The syndrome types include: exterior cold syndrome type and interior heat syndrome type. The integral vector of exterior cold syndrome type contains integral values ​​of different exterior cold dimensions. The integral vector of interior heat syndrome type contains integral values ​​of different interior heat dimensions. The syndrome type remission factor analysis module is used to obtain the initial remission factor for each syndrome type based on the uneven distribution of integral values ​​in different dimensions of the integral vector at the time of each initial sampling, combined with the patient's basic symptom information. For each syndrome type at each sampling time, the module determines the disease course influence coefficient of each syndrome type at each sampling time by the magnitude of the integral value in each dimension and the change of the integral value between previous sampling times. By adjusting the initial remission factor of another syndrome type using the disease course influence coefficient of each syndrome type at each sampling time, the module obtains the moderating remission factor of each syndrome type at each sampling time. The disease course characteristic analysis module is used to combine the regulatory and remission factors of different syndrome types at each sampling time with the integral vector at the first sampling time to obtain the Yinqiao Shuangjie effect factor of different syndrome types at each sampling time; and to combine the Yinqiao Shuangjie effect factor of two syndrome types at each sampling time with the integral vector at the sampling time to obtain the disease course risk parameters for each sampling. The evaluation model training module is used to train the model based on the disease course risk parameters and basic information data of all patients in the historical sampling database, and obtain the evaluation neural network model.

[0005] Furthermore, the method for obtaining the initial mitigation factor includes: Obtain the patient's basic disease vector and age; For any syndrome type, the symptom concentration is obtained based on the distribution entropy of the integral vector of the syndrome type at the first sampling. The symptom concentration is negatively correlated with the distribution entropy. Based on the patient's age and the magnitude of the underlying disease vector, the recovery potential of this syndrome type is obtained; combined with the magnitude of the integral vector of this syndrome type at the time of the first sampling and the recovery potential, the remission assessment value of this syndrome type is obtained. By combining the relief assessment value and symptom concentration of this syndrome type, the initial relief factor for this syndrome type is obtained.

[0006] Furthermore, the method for obtaining the recovery potential includes: The patient's age was negatively correlated and normalized to obtain the potential value of physical signs; the modulus of the patient's basic disease vector was negatively correlated to obtain the potential value of disease. The product of the physical sign potential value and the disease potential value is used as the recovery potential.

[0007] Furthermore, the method for obtaining the disease course influence coefficient includes: For any sample other than the initial sample, a certain type of evidence is used as an analysis dimension in sequence; The degree of integral change of the analysis dimension is obtained by measuring the degree of change of the integral value of the analysis dimension between the previous sampling and this sampling. By using the degree of change in the integral of the analysis dimension as a weight, the integral values ​​of the analysis dimension are weighted, summed, and normalized to obtain the disease course influence coefficient of this syndrome type.

[0008] Furthermore, the method for obtaining the integral change degree includes: The difference between the integral value of the analysis dimension at the previous sampling time and the integral value of the analysis dimension at the current sampling time is calculated and normalized to obtain the integral change of the analysis dimension.

[0009] Furthermore, the method for obtaining the regulating and mitigating factor includes: For any given syndrome type during sampling, the product of the initial remission factor of that syndrome type and the disease course influence coefficient of another syndrome type is used as the degree of regulation for that syndrome type. The sum of the degree of regulation of the syndrome and the initial relief factor of the syndrome is used as the regulation-relief factor of the syndrome.

[0010] Furthermore, the method for obtaining the active ingredient of the Yinqiao Shuangjie formula includes: For any given sample, by combining the magnitude of the integral vector of the cold syndrome type at the time of the first sample with the adjustment and relief factor at that time, the effect factor of Yinqiao Shuangjie on the cold syndrome type at that time is obtained. By combining the magnitude of the integral vector of the heat syndrome type at the time of the first sampling with the adjustment and relief factor at that time, the effect factor of Yinqiao Shuangjie on the heat syndrome type at that time of sampling is obtained. The magnitude of the integral vector of the heat syndrome type at the time of initial sampling is negatively correlated with the effect factor of Yinqiao Shuangjie.

[0011] Furthermore, the method for obtaining the disease course risk parameters includes: For any given sample, the risk parameter value at that sample is obtained by combining the modulus of the integral vector of each certificate type and the effect factor of the double solution of Yinqiao at that sample. The effect factor of the double solution of Yinqiao is negatively correlated with the risk parameter value. By combining the risk parameter values ​​of all syndrome types at the time of sampling, the disease course risk parameters for this sampling are obtained.

[0012] Furthermore, the model training based on the disease course risk parameters and basic information data of all patients in the historical sampling database to obtain the evaluation neural network model includes: Obtain the disease course risk parameters of all patients in the historical sampling database at each sampling time; use the historical sampling database as the training dataset, take the patients' disease course parameters and basic disease information as input, and take the evaluation results as output to train a deep learning model to obtain an evaluation neural network model.

[0013] Furthermore, the process of obtaining the disease course risk parameters by combining the risk parameter values ​​of all syndrome types at the time of sampling includes: The sum of the risk parameter values ​​for all syndrome types at the time of sampling is used as the disease course risk parameter for that sampling.

[0014] The present invention has the following beneficial effects: This invention collects multi-source data on different syndrome types using a cold exterior and heat interior heat syndrome model. It analyzes the initial symptoms and the mitigation factors for both cold exterior and heat interior heat syndromes. Combining this with the characteristics of syndrome type changes, it analyzes the influence between different syndrome types at different time points, adjusts mitigation factors, considers synergistic mitigation effects, and improves the accuracy of dynamic evolution analysis. Considering the mitigation effects of Yinqiao Shuangjie (a traditional Chinese medicine formula) on different syndrome types and the adjustment of mitigation factors for each syndrome type, it analyzes the patient's final disease course risk parameters. Through the time-series dynamic evolution of the disease course, a deep learning model is trained to evaluate the intervention effect of Yinqiao Shuangjie. This invention comprehensively measures disease course risk characteristics by analyzing the dynamic changes in the mitigation of different cold exterior and heat interior heat syndromes, and combines this with a deep learning model to improve the accuracy of intervention effect evaluation, providing data support for subsequent auxiliary medication decisions. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a structural diagram of a system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure, provided in one embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a Yinqiao Shuangjie compound intervention effect evaluation system based on the patient's disease course structure proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following description, in conjunction with the accompanying drawings, details the specific scheme of the Yinqiao Shuangjie Compound Intervention Effect Evaluation System provided by the present invention.

[0020] Please see Figure 1 The diagram shows a structural diagram of an intervention effect evaluation system for Yinqiao Shuangjie compound based on patient disease course structure provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a syndrome type relief factor analysis module 102, a disease course characteristic analysis module 103, and an evaluation model training module 104.

[0021] The data acquisition module 101 is used to obtain the integral vector of different syndrome types for each patient in the historical sampling database at each sampling time. The syndrome types include: exterior cold syndrome type and interior heat syndrome type. The integral vector of exterior cold syndrome type contains integral values ​​of different exterior cold dimensions. The integral vector of interior heat syndrome type contains integral values ​​of different interior heat dimensions.

[0022] According to Traditional Chinese Medicine (TCM) theory, the pathological essence of non-bacterial upper respiratory tract infections of the external cold and internal heat type is that external wind-cold invades the exterior, suppressing the defensive qi (wei yang), while the lungs and stomach already have internal heat, resulting in a combination of internal and external pathogens. Therefore, the course of the disease manifests as a dynamic interplay of two opposing physiological processes: the external cold process is characterized by wind-cold attacking the exterior, producing symptoms such as chills, lack of sweating, or headache and body aches, with more pronounced symptoms the more severe the cold. The internal heat process is characterized by excessive heat in the lungs and stomach, producing symptoms such as sore throat, dry mouth, constipation, or a red tongue with a yellow coating, with more severe symptoms the stronger the heat.

[0023] Furthermore, these two processes are not independent but rather mutually influential. If the exterior cold is not resolved, heat cannot escape; conversely, excessive interior heat accelerates the penetration of exterior cold into the interior. In other words, if the exterior cold is not eliminated, interior heat cannot dissipate, and excessive interior heat forces exterior cold further into the body. Therefore, the observation of the disease course requires simultaneous monitoring of both dynamic processes. Thus, during assessment, it is essential to independently extract quantitative characteristics for both the exterior cold syndrome and the interior heat syndrome to conduct more accurate data analysis.

[0024] In this embodiment of the invention, based on the "General Rules for the Formulation and Revision of Standards for Diagnosis and Efficacy Evaluation of Traditional Chinese Medicine Diseases" and syndrome quantification research, two sub-scales for different syndrome types were designed to collect syndrome type scores. For example, the scale for exterior cold syndrome includes the degree of aversion to cold (0–10 points), the degree of anhidrosis (0–5 points), and headache or body aches (0–5 points); the scale for interior heat syndrome includes sore throat (0–10 points), dry mouth (0–5 points), constipation (0–5 points), and red tongue with yellow coating (0–10 points). Sampling was conducted during the course of the disease treated with Yinqiao Shuangjie compound. The sampling was completed twice daily with the assistance of a doctor, specifically at 8:00 AM and 8:00 PM in one specific embodiment of the invention. The sampled data were standardized and entered into a database to obtain an integral vector. Different dimensions of the integral vector represent symptoms with different scores.

[0025] In addition, due to the possible underlying diseases of different patients, the underlying disease information and age of patients are extracted through the electronic medical record system. The underlying diseases are mainly respiratory-related diseases. The underlying disease vector is obtained by 0 and 1 labeling based on Boolean encoding. The existence is recorded as 1 and the absence is recorded as 0. The extracted data is preprocessed and standardized and then recorded in the database for subsequent data analysis.

[0026] It should be noted that the scale settings and sampling settings can be adjusted by the implementer based on the implementation scenario, and there are no restrictions here. The quantitative scoring can be standardized using Z-score. Data standardization preprocessing methods are well known to those skilled in the art and will not be elaborated here.

[0027] The syndrome type remission factor analysis module 102 is used to obtain the initial remission factor for each syndrome type based on the uneven distribution of integral values ​​in different dimensions of the integral vector at each initial sampling, combined with the patient's basic symptom information; for each syndrome type at each sampling, the course influence coefficient of each syndrome type at each sampling is determined by the magnitude of the integral value of each dimension and the change of the integral value between previous samplings; and the moderating remission factor of each syndrome type at each sampling is obtained by adjusting the initial remission factor of another syndrome type by adjusting the course influence coefficient of each syndrome type at each sampling.

[0028] To analyze the dynamic evolution of each patient's disease course, we first examined the characteristics of the initial relief of exterior cold and interior heat in each patient. The more underlying respiratory diseases a patient had, the less favorable it was for symptom relief. Furthermore, based on the clinical characteristic that a single primary symptom is easier to resolve than multiple mixed symptoms, the more concentrated the initial symptom distribution, the more focused the relief and the higher the potential for faster relief; conversely, the more dispersed the distribution, the slower the relief rate.

[0029] Therefore, to initially assess the mitigation characteristics, in this embodiment of the invention, the method for obtaining the initial mitigation factor includes: For any syndrome type, the symptom concentration of the syndrome type is obtained based on the distribution entropy of the integral vector of the syndrome type at the time of the first sampling. The symptom concentration is negatively correlated with the distribution entropy. The distribution entropy reflects the degree of distribution balance of multi-dimensional data in the vector. When the distribution entropy is larger, it indicates that the multi-dimensional symptom assessment is more balanced and the potential for relief speed is lower.

[0030] It should be noted that negative correlation represents the relationship that as one data point increases, the other decreases. The methods for negative correlation mapping can include inverse proportional values, exponential decay functions, or negative linear functions. Distribution entropy calculation and negative correlation mapping are techniques well known to those skilled in the art, and will not be limited or elaborated here.

[0031] The recovery potential is analyzed by examining the patient's physical signs. Based on the patient's age and the magnitude of the vector of underlying diseases, the recovery potential of this syndrome is obtained. Age can reflect the degree of basic individual differences in the patient's physical signs. For example, young people experience rapid penetration of external cold into the interior, while the elderly have a longer duration of external cold due to their weak constitution. At the same time, the number of underlying diseases also reflects the individual's recovery ability to a certain extent. The more underlying diseases there are, the less conducive it is to the relief of the syndrome.

[0032] Therefore, in this embodiment of the invention, the patient's age is negatively correlated and normalized to obtain the potential value of physical signs, and the magnitude of the patient's basic disease vector is negatively correlated to obtain the potential value of the disease. In a specific embodiment of the invention, the potential value of physical signs is obtained by applying an inverse proportional function to the patient's age, and the potential value of the disease is obtained by applying a negative exponential function with the natural constant as the minimum. The younger the age, the stronger the constitution may be. The smaller the magnitude of the basic disease vector, the less severe the underlying disease, which is more conducive to the dissipation of external cold pathogens or the relief of internal heat, and therefore the higher the potential.

[0033] Therefore, the product of the physical sign potential value and the disease potential value is ultimately used as the recovery potential value. The greater the recovery potential value, the higher the initial assessment of the relief ability.

[0034] Furthermore, by combining the magnitude of the integral vector and the recovery potential at the time of the initial sampling, a remission assessment value for the syndrome is obtained. A larger magnitude of the integral vector at the time of the initial sampling indicates more severe overall symptoms, suggesting a more significant effect of intervention when symptoms are severe. Conversely, a larger recovery potential indicates less resistance to remission, thus indicating a better remission trend. In this embodiment of the invention, the product of the magnitude of the integral vector and the recovery potential at the time of the initial sampling is normalized to obtain the remission assessment value.

[0035] It should be noted that normalization is a technique well known to those skilled in the art. The normalization function can be linear normalization or standard normalization, etc. The specific normalization method is not limited here.

[0036] Finally, by combining the relief assessment value and symptom concentration of the syndrome, the initial relief factor for the syndrome is obtained. The recovery potential is comprehensively assessed by the initial symptom distribution. The higher the relief assessment value and symptom concentration, the higher the potential for the relief rate and trend, and the larger the initial relief factor. In this embodiment of the invention, the product of the relief assessment value and symptom concentration of the syndrome is normalized to obtain the initial relief factor for the syndrome.

[0037] Furthermore, considering the TCM theory of mutual restraint between cold and heat, excessive external cold can constrict and obstruct the dissipation of internal heat, slowing its resolution. Conversely, excessive internal heat can force the body fluids to dissipate, potentially accelerating the transformation of external cold into internal heat, thus perpetuating or worsening the external cold. Therefore, external cold and internal heat sometimes exhibit a synergistic effect. After the external cold is relieved, the body's heat can be released more easily, and the flow of qi and blood becomes smoother, thus improving the symptoms of internal heat. Similarly, after the internal heat is relieved, the accumulation of heat in the body decreases, the binding force of the cold weakens, and the symptoms of external cold will also improve.

[0038] Therefore, by analyzing the temporal changes and numerical values ​​of multidimensional symptoms during the course of the disease, the coefficient that can influence another syndrome type is determined. Preferably, in this embodiment of the invention, the method for obtaining the disease course influence coefficient includes: For any sample of a syndrome type other than the initial sample, each dimension of that syndrome type is used as an analysis dimension in sequence, and the symptom manifestations reflected by each dimension under a single syndrome type are analyzed sequentially. Then, by measuring the degree of change in the integral value of the analysis dimension between the previous sample and the current sample, the integral change degree of the analysis dimension is obtained. The higher the degree of change, the better the relief effect on the corresponding symptoms.

[0039] In this embodiment of the invention, the difference between the integral value of the analysis dimension at the previous sampling time and the integral value of the analysis dimension at the current sampling time is calculated and normalized to obtain the integral change degree of the analysis dimension. The relief performance under the interference course is measured by the change of the integral of a single symptom. The higher the integral change degree, the more significant the relief performance.

[0040] For symptoms with low scores, their ability to reflect the intervention effect is weak, and the impact on relief is considered from the perspective of participation. However, when the score is high, it can significantly reflect the intervention effect. Therefore, the influence coefficient is calculated by combining the changes in scores across different dimensions. Thus, the degree of change in the scores of the analytical dimensions is used as a weight. The scores of the analytical dimensions are weighted, summed, and normalized to obtain the course influence coefficient for this syndrome type. This is achieved by multiplying the degree of change in the scores of the analytical dimensions by the scores of the analytical dimensions, and then normalizing the sum of the products after multiplying all dimensions.

[0041] A higher disease course influence coefficient indicates a more significant remission of this syndrome type at the time of sampling, which greatly promotes the remission ability of another syndrome type. Therefore, in evaluating the intervention effect on disease course, the remission factor is adjusted based on the synergistic effect. In this embodiment of the invention, the method for obtaining the remission factor includes: For any given symptom type at any sampling time, the product of the initial remission factor of the symptom type and the disease course influence coefficient of another symptom type is used as the degree of regulation of the symptom type. The sum of the degree of regulation of the symptom type and the initial remission factor of the symptom type is used as the moderating remission factor of the symptom type. The higher the influence coefficient, the greater the remission assessment value that can be represented at the corresponding sampling time.

[0042] This completes the assessment of the relief of each syndrome type at each sampling point in the timeline of the patient's intervention and treatment course.

[0043] The disease course characteristic analysis module 103 is used to combine the regulatory and remission factors of different syndrome types at each sampling time with the integral vector at the first sampling time to obtain the Yinqiao Shuangjie effect factor of different syndrome types at each sampling time; and to combine the Yinqiao Shuangjie effect factor of two syndrome types at each sampling time with the integral vector at the sampling time to obtain the disease course risk parameters for each sampling.

[0044] By analyzing the moderating and alleviating factors in the exterior cold and interior heat syndromes during time-series upsampling, the evolution of syndrome patterns during the intervention course can be reflected. However, the severity of different initial syndrome patterns also has a certain impact on the dynamic evolution of alleviation. For example, in patients with severe initial exterior cold syndrome, the exterior-releasing effect of Yinqiao Shuangjie is relatively rapid, while in patients with severe initial interior heat syndrome, the heat-clearing effect of Yinqiao Shuangjie is slow to take effect, and symptom relief is relatively slow. Therefore, the effect of Yinqiao Shuangjie at the time of sampling is evaluated for different alleviation manifestations. In this embodiment, the method for obtaining the Yinqiao Shuangjie effect factors includes: For any given sample, the modulus of the integral vector of the cold syndrome at the time of the first sample and the adjustment and relief factor at that time are combined to obtain the Yinqiao Shuangjie effect factor of the cold syndrome at that time. In this embodiment of the invention, the product of the modulus of the integral vector of the cold syndrome at the time of the first sample and the adjustment and relief factor at that time is normalized to obtain the Yinqiao Shuangjie effect factor of the cold syndrome. The more severe the initial cold syndrome, the stronger the relieving effect of Yinqiao Shuangjie.

[0045] Furthermore, by combining the magnitude of the integral vector of the heat syndrome at the initial sampling time with the regulating and alleviating factor at that sampling time, the effect factor of Yinqiao Shuangjie on the heat syndrome at that sampling time is obtained. The magnitude of the integral vector of the heat syndrome at the initial sampling time is negatively correlated with the effect factor of Yinqiao Shuangjie. In this embodiment of the invention, the ratio between the regulating and alleviating factor at that sampling time and the magnitude of the integral vector of the heat syndrome at the initial sampling time is normalized to obtain the effect factor of Yinqiao Shuangjie on the heat syndrome at that sampling time. The lower the severity of the initial heat syndrome, the stronger the heat-clearing effect of Yinqiao Shuangjie.

[0046] The stronger the effect of Yinqiao Shuangjie, the more significant the symptom relief and the lower the risk during the course of the disease. Therefore, by combining the integral vector at the time-series upsampling with the Yinqiao Shuangjie effect factor, the risk level of the disease course at this time can be comprehensively reflected, which is convenient for subsequent evaluation of the intervention effect.

[0047] In this embodiment of the invention, for any sample, the risk parameter value at that sample is obtained by combining the magnitude of the integral vector of each certificate type at that sample with the effect factor of the double solution of the silver and the qiao. The effect factor of the double solution of the silver and the risk parameter value are negatively correlated. In this embodiment of the invention, the risk parameter value is obtained by multiplying the value of the effect factor of the double solution of the silver and the qiao at that sample by the negative correlation mapping of the double solution of the silver and the magnitude of the integral vector of each certificate type at that sample. The smaller the effect factor of the double solution of the silver and the larger the magnitude of the integral vector, the higher the risk.

[0048] Finally, by combining the risk parameter values ​​of all syndrome types at the time of sampling, the disease course risk parameter of the sampling is obtained. In this embodiment of the invention, the sum of the risk parameter values ​​of all syndrome types at the time of sampling is used as the disease course risk parameter of the sampling. The syndrome types reflect the disease course parameters and are used for subsequent training analysis.

[0049] The evaluation model training module 104 is used to train the model based on the disease course risk parameters and basic information data of all patients in the historical sampling database to obtain the evaluation neural network model.

[0050] Using a historical database as a training set, a deep learning model is employed to uncover the evolutionary correlation between changes in time-series disease progression and intervention effects. In this embodiment of the invention, disease progression risk parameters for all patients at each sampling time, as well as efficacy indicators after evaluation, are obtained from the historical sampling database. Using the historical sampling database as the training dataset, the labeled dataset can be divided into a training set, a validation set, and a test set in a 7:2:1 ratio after analysis.

[0051] In one specific embodiment of this invention, an LSTM+Dense+Linear deep learning model is selected, with the mean squared error function as the loss function. The Adam optimizer is chosen, and the initial learning rate is set to 0.001. A learning rate decay strategy is used to avoid later oscillations. The model input consists of time-sampled disease course risk parameters and basic patient information, including age and basic symptom vectors. The model outputs a single continuous value, i.e., the efficacy index, resulting in a trained evaluation neural network model. This evaluation neural network model can then be used to assess new patients and assist in medication decisions. It should be noted that the specific training process is a technique well-known to those skilled in the art and will not be elaborated further here.

[0052] In summary, this invention collects multi-source data on different syndrome types using a cold exterior and heat interior heat syndrome table. It analyzes the initial symptom profile to identify allergens in patients with cold exterior and heat interior heat syndromes, and combines this with an analysis of syndrome type changes to examine the influence between different syndrome types at different time points. By adjusting allergens and considering synergistic allergen effects, the accuracy of dynamic evolution analysis is improved. Considering the allergic effects of Yinqiao Shuangjie (a traditional Chinese medicine formula) on different syndrome types and the adjustment of allergens based on syndrome types, the final disease course risk parameters of patients are analyzed. A deep learning model is trained based on the temporal dynamic evolution of the disease course to evaluate the intervention effect of Yinqiao Shuangjie. This invention comprehensively measures disease course risk characteristics through the dynamic changes in the allergy of different cold exterior and heat interior heat syndromes, and, combined with a deep learning model, improves the accuracy of intervention effect evaluation, providing data support for subsequent auxiliary medication decisions.

[0053] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0054] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure, characterized in that, The system includes: The data acquisition module is used to obtain the integral vector of different syndrome types for each patient in the historical sampling database at each sampling time. The syndrome types include: exterior cold syndrome type and interior heat syndrome type. The integral vector of exterior cold syndrome type contains integral values ​​of different exterior cold dimensions. The integral vector of interior heat syndrome type contains integral values ​​of different interior heat dimensions. The syndrome type remission factor analysis module is used to obtain the initial remission factor for each syndrome type based on the uneven distribution of integral values ​​in different dimensions of the integral vector at the time of each initial sampling, combined with the patient's basic disease information. For each syndrome type at each sampling time, the module determines the disease course influence coefficient for each syndrome type at each sampling time by the magnitude of the integral value in each dimension and the change of the integral value between previous sampling times. By adjusting the initial remission factor of another syndrome type using the disease course influence coefficient of each syndrome type at each sampling time, the module obtains the moderating remission factor for each syndrome type at each sampling time. The disease course characteristic analysis module is used to combine the regulatory and remission factors of different syndrome types at each sampling time with the integral vector at the first sampling time to obtain the Yinqiao Shuangjie effect factor of different syndrome types at each sampling time; and to combine the Yinqiao Shuangjie effect factor of two syndrome types at each sampling time with the integral vector at the sampling time to obtain the disease course risk parameters for each sampling. The evaluation model training module is used to train the model based on the disease course risk parameters and basic information data of all patients in the historical sampling database, and obtain the evaluation neural network model.

2. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The method for obtaining the initial mitigation factor includes: Obtain the patient's basic disease vector and age; For any syndrome type, the symptom concentration is obtained based on the distribution entropy of the integral vector of the syndrome type at the time of the first sampling. The symptom concentration is negatively correlated with the distribution entropy. Based on the patient's age and the magnitude of the underlying disease vector, the recovery potential of this syndrome type is obtained; combined with the magnitude of the integral vector of this syndrome type at the time of the first sampling and the recovery potential, the remission assessment value of this syndrome type is obtained. By combining the relief assessment value and symptom concentration of this syndrome type, the initial relief factor for this syndrome type is obtained.

3. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 2, characterized in that, The method for obtaining the recovery potential includes: The patient's age was negatively correlated and normalized to obtain the potential value of physical signs; the modulus of the patient's basic disease vector was negatively correlated to obtain the potential value of disease. The product of the physical sign potential value and the disease potential value is used as the recovery potential.

4. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The methods for obtaining the disease course influence coefficient include: For any sample other than the initial sample, a certain type of evidence is used as an analysis dimension in sequence; The degree of integral change of the analysis dimension is obtained by measuring the degree of change in the integral value of the analysis dimension between the previous sampling and this sampling. By using the degree of change of the integral of the analysis dimension as the weight, the integral values ​​of the analysis dimension are weighted, summed, and normalized to obtain the disease course influence coefficient of this syndrome type.

5. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 4, characterized in that, The method for obtaining the integral change degree includes: The difference between the integral value of the analysis dimension at the previous sampling time and the integral value of the analysis dimension at the current sampling time is calculated and normalized to obtain the integral change of the analysis dimension.

6. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The method for obtaining the regulatory and mitigating factors includes: For any given syndrome type during sampling, the product of the initial remission factor of that syndrome type and the disease course influence coefficient of another syndrome type is used as the degree of regulation for that syndrome type. The sum of the degree of regulation of the syndrome and the initial relief factor of the syndrome is used as the regulation-relief factor of the syndrome.

7. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The method for obtaining the active ingredient of Yinqiao Shuangjie includes: For any given sample, by combining the magnitude of the integral vector of the cold syndrome type at the time of the first sample with the adjustment and relief factor at that time, the effect factor of Yinqiao Shuangjie on the cold syndrome type at that time is obtained. By combining the magnitude of the integral vector of the heat syndrome type at the time of the first sampling with the adjustment and relief factor at that time, the effect factor of Yinqiao Shuangjie on the heat syndrome type at that time of sampling is obtained. The magnitude of the integral vector of the heat syndrome type at the time of initial sampling is negatively correlated with the effect factor of Yinqiao Shuangjie.

8. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The methods for obtaining the disease course risk parameters include: For any given sample, the risk parameter value at that sample is obtained by combining the modulus of the integral vector of each certificate type and the effect factor of the double solution of Yinqiao at that sample. The effect factor of the double solution of Yinqiao is negatively correlated with the risk parameter value. By combining the risk parameter values ​​of all syndrome types at the time of sampling, the disease course risk parameters for this sampling are obtained.

9. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 1, characterized in that, The model is trained based on the disease course risk parameters and basic information data of all patients in the historical sampling database to obtain an evaluation neural network model, including: Obtain the disease course risk parameters of all patients in the historical sampling database at each sampling time; use the historical sampling database as the training dataset, take the patients' disease course parameters and basic disease information as input, and take the evaluation results as output to train a deep learning model to obtain an evaluation neural network model.

10. The system for evaluating the intervention effect of Yinqiao Shuangjie compound based on the patient's disease course structure according to claim 8, characterized in that, The process risk parameters for this sampling are obtained by combining the risk parameter values ​​of all syndrome types at the time of sampling, including: The sum of the risk parameter values ​​for all syndrome types at the time of sampling is used as the disease course risk parameter for that sampling.

Images